Hi Daniel, this is how ActiveMQ Stomp implementation works. Take a look at:
http://activemq.apache.org/stomp.html#Stomp-WorkingwithJMSText/BytesMessagesandStomp if you send a content-length header, it will be treated as byte message. I don't see how it could be changed between 5.1 and 5.2, because it works this way for quite a long time. As for the content-type, protocol (http://stomp.codehaus.org/Protocol) does not specify anything about it, so it is ignored. I'm not sure why your client library set this value. However, we plan to support better message transformation for Stomp 1.1 (which is already implemented in ActiveMQ). Take a look at the following links for more info http://activemq.apache.org/stomp.html#Stomp-Messagetransformations http://stomp.codehaus.org/Stomp+v1.1+Ideas#Stompv1.1Ideas-Messagetransformation Cheers -- Dejan Bosanac http://www.ttmsolutions.com - get a free ActiveMQ user guide ActiveMQ in Action - http://www.manning.com/snyder/ Scripting in Java - http://www.scriptinginjava.net Daniel Jabbour wrote: > Team- > > I recently upgraded from ActiveMQ 5.1 to ActiveMQ 5.2 broker. I am > sending messages from a Ruby STOMP client and consuming them in a JMS > application. Previously, the messages were coming through as > TextMessages (with ActiveMQ 5.1). After the upgrade, I started getting > ActiveMQBytesMessages. I noticed some list postings back in 2006 about > this issue... but it's interesting it resurfaced after my ActiveMQ > broker upgrade (with no code changes to my producer or consumers). > > The summary of the issue as I understand it: > * The Ruby STOMP client sends "content-length" headers. > * Previously, it was suggested to patch the Ruby STOMP client to omit > this header, thus forcing JMS/ActiveMQ to evaluate the message as a > TextMessage. > * However, the STOMP client is also sending "content-type: > text/plain; charset=UTF-8" in the headers. > > So my question is two-fold: > * What changes from ActiveMQ 5.1 to 5.2 that broke the STOMP client's > ability to send TextMessages? > * It seems silly to patch the STOMP client to omit a header to force > the message type. Why is ActiveMQ/JMS ignoring the content-type: > text/plain header? Wouldn't it make more sense to evaluate that header > to determine the message type? > > Thank you very much in advance, > > Daniel > > -- > Daniel Jabbour > Executive Vice President of Systems Architecture > Office: (646) 778-3789 x707 > [EMAIL PROTECTED] > InfoExchange, Inc. > >
